Objective 2.1

PMLE

Exploring and preprocessing data for ML

Objective 2.1 sits in Collaborating within and across teams to manage data and models, which carries 16% of the ML Engineer exam. The questions below are original, written from the official objective title above, and each explanation cites the Google Cloud page it rests on.

Objective title verbatim from the official objectives. Google Cloud exam page

A worked example

Shown solved, with the whole explanation open: this is what every question here carries.

2-1Collaborating within and across teams to manage data and models

A team already has feature tables in BigQuery and wants online serving without creating catalog entries first. Is registration required before serving?

Yes, every feature column must be registered firstNothing forces a column through the catalog.
No, features can be served without registering themCorrect · your answerCorrect. A view can point straight at the table.
Yes, but only for tables carrying a timestampTimestamp handling is a reason to register, not a rule.
No, though serving then runs in batch mode onlyReal-time reads do not depend on that step.

Correct.

Concept

A registry is a catalogue laid over data that already lives somewhere else. Making it mandatory would turn a convenience into a gate, so it stays optional and earns its keep only where the extra bookkeeping is wanted.

Why B

Registering features is optional. Features can be served online even when the BigQuery data sources are not added to the Feature Registry, though registration brings timestamp handling, column selection, aggregation across sources and drift monitoring.

Source

Registering your features is optional. You can serve features online even if you don't add your BigQuery data sources to the Feature Registry. However, registering your features is advantageous in the following scenarios: Your data contains multiple instances of the same entity ID and you need to prepare your data in a time-series format with a timestamp column. When you register your features, Agent Platform Feature Store looks up the timestamp and serves only the latest feature values. You want to register specific feature columns from a data source. You want to aggregate specific columns…

About Feature Store, Agent Platform documentation, checked August 2026
#gcp#feature-store#mlops#data-management

Now you: objective 2.1 questions

No account needed. The explanation opens when you answer.

Sample question 1 of 3

2-1Collaborating within and across teams to manage data and models

An engineer has several rows per entity ID in a source table and needs the freshest values served. What does registering those features add?

Sample question 2 of 3

2-1Collaborating within and across teams to manage data and models

A team creates an online serving cluster and cannot find where to name the BigQuery source on it. Which resource carries the data source?

Sample question 3 of 3

2-1Collaborating within and across teams to manage data and models

A team must serve terabytes of frequently updated features online and has no embeddings to serve. Which online serving type fits that shape?

Full ML Engineer question bank coming

We’re writing the complete bank from the official objectives right now. Leave your email and we’ll tell you when it ships, nothing else, ever.

Read the sources

These are the official pages the questions above cite. Reading them is studying the objective from the primary source, which is what the explanations point you toward anyway.

More objectives in Collaborating within and across teams to manage data and models